Cyanate Ester/Polyhedral Oligomeric Silsesquioxane (POSS) Nanocomposites: Synthesis and Characterization

Cyanate Ester/Polyhedral Oligomeric Silsesquioxane (POSS)Nanocomposites: Synthesis and Characterization†

Kaiwen Liang,‡ Guizhi Li,§ Hossein Toghiani,‡ Joseph H. Koo,| andCharles U. Pittman, Jr.*,§

DaVe C. Swalm School of Chemical Engineering, Mississippi State UniVersity, Mississippi State,Mississippi 39762, Department of Chemistry, Mississippi State UniVersity, Mississippi State, Mississippi

39762, and Department of Mechanical Engineering, Texas A&M UniVersity, College Station, Texas 78712

ReceiVed July 19, 2005

Cyanate ester (PT-15, Lonza Corp.) composites containing the blended polyhedral oligomericsilsesquioxane (POSS), TriSilanolPhenyl-POSS (C42H38O12Si7), were prepared containing PT-15/POSS99/1, 97/3, 95/5, 90/10, and 85/15 w/w ratios. The composites were characterized by FT-TR, X-raydiffraction (XRD), small-angle neutron scattering (SANS), scanning electron microscopy (SEM), X-rayenergy dispersive spectroscopy (X-EDS), transmission electron microscopy (TEM), dynamic mechanicalthermal analysis (DMTA), and three-point bending flexural tests. TriSilanolPhenyl-POSS was throughlydispersed into uncured liquid PT-15 resin. After curing, XRD, SANS, and X-EDS measurements wereconsistent with partial molecular dispersion of a portion of the POSS units in the continuous matrixphase while the remainder forms POSS aggregates. Larger aggregates are formed at higher loadings.SANS, SEM, and TEM show that POSS-enriched nanoparticles are present in the PT-15/POSS composites.The storage bending moduli,E′, and the glass transition temperatures,Tg, of PT-15/POSS 99/1, 97/3,and 95/5 composites are higher than those of the pure PT-15 over the temperature range from 35 to 350°C. The E′ values for all these composites (except for the 15 wt % POSS sample) are significantlygreater than that of the pure resin atT > Tg. Therefore, small amounts (e5 wt %) of TriSilanolPhenyl-POSS incorporated into cyanate ester resin PT-15 can improve the storage modulus and the high-temperature properties of these cyanate ester composites versus the pure PT-15 resin. The flexural strengthand flexural modulus are also raised by POSS incorporation.

Introduction

The development of hybrid organic polymer-inorganicnanocomposites with a variety of new and improved proper-ties has attracted much research interest in the past fewyears.1 The inorganic nanophases being studied include nano-clays,2 carbon nanotubes,3 vapor-grown carbon nanofibers,4

various inorganic nanofibers,5 and polyhedral oligomeric sil-sesquioxanes.1s-z,6 Silsesquioxane is the term for all structureswith the formula (RSiO1.5)n, where R is hydrogen or anyalkyl, alkylene, aryl, arylene, or organofunctional derivativeof alkyl, alkylene, aryl, or arylene groups. Silsesquioxanesmay form ladder,7 cage,7a,8partial cage,9 and polymer struc-tures.1z,10Among various types of silsesquioxanes, polyhedraloligomeric silsesquioxanes (POSS) reagents offer a uniqueopportunity for preparing hybrid organic-inorganic materialswith the inorganic structural units truly molecularly dispersedwithin the nanocomposites. Typical POSS cages have theempirical formulas (RSiO1.5)8,10,or12. These are referred to asT8, T10, and T12 cages, respectively. Each cage silicon atomis attached to a single R substituent and these substituentscan be organic (cyclohexyl, phenyl, etc.) or inorganic-

organic hybrids (e.g.,-OSiMe2OPh). Incompletely closedcage structures are also possible. Two sample POSS struc-tures, both a T8 cage and an incompletely closed cagemolecule1 are shown below. POSS nanostructured chemi-cals, with sizes from 1 to 3 nm in diameter, can be thoughtof as the smallest possible particles of silica. However, un-like silica, silicones, or fillers, each POSS molecule con-tains either nonreactive or reactive organic substituents atthe corner silicon atoms. These organic substituents canmake these POSS molecules compatible with polymers ormonomers.

POSS derivatives are now available with reactive func-tionalities suitable for polymerization or grafting.1s-z,6,11

Hence, POSS nanostructured chemicals can be incorporatedinto common plastics via copolymerization,11c,12grafting,11c,13

or blending,11c,13,14thereby offering a special opportunity forthe preparation of new thermoset6,15 and thermoplastic

† This work is taken from the Ph.D. dissertation of Kaiwen Liang.* Corresponding author. E-mail: [email protected].‡ Dave C. Swalm School of Chemical Engineering, Mississippi State

University.§ Department of Chemistry, Mississippi State University.| Texas A&M University, current address: Department of Mechanical

Engineering, University of Texas at Austin, Austin, TX 78712.

301Chem. Mater.2006,18, 301-312

10.1021/cm051582s CCC: $33.50 © 2006 American Chemical SocietyPublished on Web 12/18/2005

materials.1v,z,12b,c,13,14,16The incorporation of POSS derivativesinto polymeric materials can lead to substantial improvementsin polymer properties including increases in use tempera-ture,17 oxidation resistance,13 surface hardening,13 and me-chanical properties,18 as well as reductions in flammability,19

heat evolution,20 and viscosity21 during processing. Theseimprovements have been shown to apply to a wide range ofthermoplastics and thermoset systems, i.e., methacrylates,22

styrenes,23 norbornenes,24 ethylenes,25 epoxies,26 siloxanes,12c

etc.Cyanate esters are currently in widespread use because of

their high thermal stability, excellent mechanical propertiesflame resistance, low outgassing, and radiation resistance.27

Applications of cyanate ester include structural aerospace,electronic, microwave-transparent composites, encapsulants,and adhesives. Cyanate ester resins are superior to conven-tional epoxy, polyimide, and BMI resins. For example, themoisture absorption rate of cyanate esters is lower than thoseof epoxy, polyimide, and BMI resins. High glass transitiontemperatures (>250°C) of cyanate esters fill a temperature

(1) (a) Giannelis, E. P.AdV. Mater.1996, 8, 29-35. (b) Jog, J. P.; Hambir,S.; Bulakh, N.Polym. Eng. Sci.2002, 42, 1800-1807. (c) Kim, D.S.; Lee, K. M.J. Appl. Polym. Sci.2003, 90, 2629-2633. (d) Liu, Z.J.; Chen, K. Q.; Yan, D. Y.Eur. Polym. J.2003, 39, 2359-2366. (e)Priya, L.; Jog, J. P.J. Polym. Sci., Part B: Polym. Phys.2002, 40,1682-1689. (f) Wang, Y. Z.; Zhang, L. Q.; Tang, C. H.; Yu, D. S.J.Appl. Polym. Sci.2000, 78, 1879-1883. (g) Bureau, M. N.; Denault,J.; Cole, K. C.; Enright, G. D.Polym. Eng. Sci.2002, 42, 1897-1906. (h) Wan, C. Y.; Qiao, X. Y.; Zhang, Y.; Zhang, Y. X.Polym.Test.2003, 22, 453-461. (i) Wu, S. H.; Wang, F. Y.; Ma, C. C. M.;Chang, W. C.; Kuo, C. T.; Kuan, H. C.; Chen, W. J.Mater. Lett.2001, 49, 327-333. (j) Burnside, S. D.; Giannelis, E. P. Chem. Mater.1995, 7, 1597-1600. (k) Agag, T.; Koga, T.; Takeichi, T.Polymer2001, 42, 3399-3408. (l) Kodgire, P.; Kalgaonkar, R.; Hambir, S.;Bulakh, N.; Jog, J. P.J. Appl. Polym. Sci.2001, 81, 1786-1792. (m)Chen, Z.; Gong, K.J. Appl. Polym. Sci.2002, 84, 1499-1503. (n)Gilman, J. W.Appl. Clay. Sci.1999, 15, 31-49. (o) Lepoittevin, B.;Devalckenaere, M.; Pantoustier, N.; Alexandre, M.; Kubies, D.;Calberg, C.; Jerome, R.; Dubois, P.Polymer2002, 43, 4017-4023.(p) Gu, A. J.; Liang, G. Z.Polym. Degrad. Stab.2003, 80, 383-391.(q) Bourbigot, S.; Devaux, E.; Flambard, X.Polym. Degrad. Stab.2002, 75, 397-402. (r) Gilman, J. W.; Kashiwagi, T.; Lichtenhan, J.D. SAMPE J.1997, 33, 40-46. (s) Romo-Uribe, A.; Mather, P. T.;Haddad, T. S.; Lichtenhan, J. D.J. Polym. Sci., Part B: Polym. Phys.1998, 36, 1857-1872. (t) Mather, P. T.; Jeon, H. G.; Romo-Uribe,A.; Haddad, T. S.; Lichtenhan, J. D.Macromolecules1999, 32, 1194-1203. (u) Lee, A.; Lichtenhan, J. D.J. Appl. Polym. Sci.1999, 73,1993-2001. (v) Jeon, H. G.; Mather, P. T.; Haddad, T. S.Polym. Int.2000, 49, 453-457. (w) Fu, B. X.; Hsiao, B. S.; Pagola, S.; Stephens,P.; White, H.; Rafailovich, M.; Sokolov, J.; Mather, P. T.; Jeon, H.G.; Phillips, S.; Lichtenhan, J. D.; Schwab, J.Polymer2001, 42, 599-611. (x) Gonzalez, R. I.; Phillips, S. H.; Hoflund, G. B.J. Spacecr.Rockets2000, 37, 463-467. (y) Lichtenhan, J. D.; Otonari, Y. A.;Carr, M. J.Macromolecules1995, 28, 8435-8437. (z) Haddad, T.S.; Lichtenhan, J. D.Macromolecules1996, 29, 7302-7304.

(2) (a) Simon, S. L.; Gillham, J. K.J. Appl. Polym. Sci.1993, 47, 461-485. (b) Yeh, J. M.; Chin, C. P.J. Appl. Polym. Sci.2003, 88, 1072-1080. (c) Wang, Z.; Pinnavaia, T. J.Chem. Mater.1998, 10, 1820-1826. (d) Alexandre, M.; Dubois, P.Mater. Sci. Eng. R Rep.2000,28, 1-63. (e) Gangun, S.; Dean, D.Polymer2003, 44, 1315-1319.(f) Ganquli, S.; Dean, D.; Vaia, R.Polym. Mater. Sci. Eng.2002, 87,94-96.

(3) (a) Dufresne, A.; Paillet, M.; Putaux, J. L.; Canet, R.; Carmona, F.;Delhaes, P.J. Mater. Sci.2002, 37, 3915-3923. (b) Penumadu, D.;Dutta, A.; Pharr, G. M.; Files, B.J. Mater. Res.2003, 18, 1849-1853. (c) Thostenson, E. T.; Chou, T. W.J. Phys. D: Appl. Phys.2002, 35, L77-L80. (d) Bai, J. B.; Allaoui, A.Compos. Part A: Appl.Sci. Manuf.2003, 34, 689-694.

(4) (a) Kuriger, R. J.; Alam, M. K.; Anderson, D. P.; Jacobsen, R. L.Compos. Part A: Appl. Sci. Manuf.2002, 33, 53-62. (b) Shofner,M. L.; Rodriguez-Macias, F. J.; Vaidyanathan, R.; Barrera, E. V.Compos. Part A: Appl. Sci. Manuf.2003, 34, 1207-1217. (c) Kuriger,R. J.; Alam, M. K.Polym. Compos.2001, 22, 604-612. (d) Gordeyev,S. A.; Ferreira, J. A.; Bernardo, C. A.; Ward, I. M.Mater. Lett.2001,51, 32-36. (e) Lakshminarayanan, P. V.; Toghiani, H.; Pittman, C.U., Jr. Carbon2004, 42, 2433-2442. (f) Patton, R. D.; Pittman, C.U., Jr.; Wang, L.Int. SAMPE Tech. Conf.1997, 29, 77-84. (g) Patton,R. D.; Pittman, C. U., Jr.; Wang, L.; Hill, J. R.; Day, A.Compos.Part A: Appl. Sci. Manuf.2001, 33, 243-251.

(5) (a) Yamamoto, K.; Otsuka, H.; Takahara, A.; Wada, S. I.J. Adhes.2002, 78, 591-602. (b) Xu, H. Y.; Kong, H.; Yang, Z. B.Chin. J.Mater. Res.2003, 17, 127-131.

(6) Lee, A.; Lichtenhan, J. D.Macromolecules1998, 31, 4970-4974.(7) (a) Unno, M.; Suto, A.; Takada, K.; Matsumoto, H.Bull. Chem. Soc.

Jpn.2000, 73, 215-220. (b) Xie, P.; Zhang, R.Polym. AdV. Technol.1997, 8, 649-656. (c) Xu, H.; Xie, P.; Zhang, R.Eur. Polym. J.2001,37, 2397-2405.

(8) (a) Lichtenhan, J. D.Comments Inorg. Chem.1995, 17, 115-130.(b) Lichtenhan, J. D.Polymeric Material Encyclopedia; Salamone, J.C., Ed.; CRC Press: New York, 1996; Vol. 10, pp 7768-7778. (c)Feher, F. J.; Budzichowski, T. A.Polyhedron1995, 14, 3239-3253.

(9) Deng, J.; Polidan, J. T.; Hottle, J. R.; Farmer-Creely, C. E.; Viers, B.D.; Esker, A. R.J. Am. Chem. Soc.2002, 124, 15194-15195.

(10) Pyun, J.; Matyjaszewski, K.; Wu, J.; Kim, G. M.; Chun, S. B.; Mather,P. T. Polymer2003, 44, 2739-2750.

(11) (a) Schwab, J. J.; Lichtenhan, J. D.; Chaffee, K. P.; Mather, P. T.;Romo-Uribe, A.Mater. Res. Soc. Symp. Proc.1998, 519, 21-27. (b)Klein, L. C., Francis, L. F., DeGuire, M. R., Mark, J. E., Eds.Organic/inorganic hybrid materials II. MRS Symposium Proceedings, 1999; p576. (c) Lichtenhan, J. D.; Schwab, J. J.Int. SAMPE Tech. Conf.2000,32, 185-191. (d) Li, G.; Wang, L.; Ni, H.; Pittman, C. U., Jr.J. Inorg.Organomet. Polym.2001, 11, 123.

(12) (a) Mather, P. T.; Jeon, H. G.; Haddad, T. S.Polym. Prepr.2000, 41,528-529. (b) Haddad, T. S.; Choe, E.; Lichtenhan, J. D.Mater. Res.Soc. Symp. Proc.1996, 435, 25-32. (c) Haddad, T. S.; Stapleton, R.;Jeon, H. G.; Mather, P. T.; Lichtenhan, J. D.; Phillips, S. H.Polym.Prepr. 1999, 40, 496-497.

(13) Phillips, S. H.; Blanski, R. L.; Svejda, S. A.; Haddad, T. S.; Lee, A.;Lichtenhan, J. D.; Feher, F. J.; Mather, P. T.; Hsiao, B. S.Mater.Res. Soc. Symp. Proc.2000, 628, CC4.6.1-CC4.6.10.

(14) Blanski, R. L.; Phillips, S. H.; Chaffee, K.; Lichtenhan, J. D.; Lee,A.; Geng, H. P.Polym. Prepr.2000, 41, 585-586.

(15) (a) Haddad, T. S.; Lee, A.; Mather, P. T.; Phillips, S. H.Polym. Prepr.2000, 41, 584. (b) Lu, S.; Martin, G. C.Conf. Proc./ANTEC2003, 2,1893-1897. (c) Constable, G. S.; Coughlin, E. B.; Lesser, A. J.Conf.Proc./ANTEC2003, 2, 1663-1667. (d) Li, G.; Wang, L.; Toghiani,H.; Pittman, C. U., Jr.; Daulton, T. L.; Koyama, K.Macromolecules2001, 34, 8686-8693. (e) Li, G.; Wang, L.; Toghiani, H.; Daulton,T. L.; Pittman, C. U., Jr.Polymer2002, 43, 4167-4176. (f) Gupta,S. K.; Schwab, J. J.; Lee, A.; Fu, B. X.; Hsiao, B. S.Int. SAMPESymp. Exhib.2002, 47, 1517-1526. (g) Feher, F. J.; Lucke, S.;Schwab, J. J.; Lichtenhan, J. D.; Phillips, S. H.; Lee, A.Polym. Prepr.2000, 41, 526.

(16) (a) Haddad, T. S.; Mather, P. T.; Jeon, H. G.; Romo-Uribe, A.; Farris,A. R.; Lichtenhan, J. D.Mater. Res. Soc. Symp.-Proc. V 519, Org./Inorg. Hybrid Mater.1998, 381-386. (b) Mantz, R. A.; Jones, P. F.;Chaffee, K. P.; Lichtenhan, J. D.; Ismail, M. K.; Burmeister, M.Chem.Mater. 1996, 8, 1250-1259.

(17) (a) Xu, H. Y.; Kuo, S. W.; Lee, J. Y.; Chang, F. C.Polymer2002,43, 5117-5124. (b) Horwath, J.; Schweickart, D.IEEE 2002, 644-647.

(18) (a) Pellice, S. A.; Fasce, D. P.; Williams, R. J. J.J. Polym. Sci., PartB: Polym. Phys.2003, 41, 1451-1461. (b) Lee, A.Mater. Res. Soc.Symp.-Proc. 1999, 576, 343-350.

(19) Philips, S. H.; Gonzalez, R. I.; Chaffee, K. P.; Haddad, T. S.; Hoflund,G. B.; Hsiao, B. S.; Fu, B. X.SAMPE2000, 45, 1921-1932.

(20) Huang, J. C.; He, C. B.; Xiao, Y.; Mya, K. Y.; Dai, J.; Siow, Y. P.Polymer2003, 44, 4491-4499.

(21) Fu, B. X.; Namani, M.; Lee, A.Polymer2003, 44, 7739-7747.(22) Lichtenhan, J. D.; Otonari, Y. A.; Carr, M. J.Macromolecules1995,

28, 8435.(23) Haddad, T. S.; Lichtenhan, J. D.Macromolecules1996, 29, 7302-

7304.(24) Bharadwaj, R. K.; Berry, R. J.; Farmer, B. L.Polymer2000, 41, 7209-

7221.(25) Tsuchida, A.; Bolln, C.; Sernetz, F. G.; Frey, H.; Mulhaupt, R.

Macromolecules1997, 30, 2818.(26) Lee, A.; Lichtenhan, J. D.Macromolecules1998, 31, 4970-4974.(27) (a) Simon, S. L.; Gillham, J. K.J. Appl. Polym. Sci.1993, 47, 461.

(b) Brand, R. A.; Harrison, E. S.NASA CR1982, 50. (c) Das, S.;Prevorsek, D. C.; Debona, B. T.21st SAMPE Tech. Conf.1989, 21,972. (d) McConnell, V. P.AdV. Compos.1992, 7, 28. (e) Shimp, D.A.; Christenson, J. R.; Ising, S. J.34th Int. SAMPE Symp.1991, 36,336. (f) Snow, A. W.; Buckley, L. J.; Armistead, J. P.J. Polym. Sci.,Polym. Chem.1999, 37, 135-150. (g) Hamerton, I.Chemistry andTechnology of Cyanate Ester Resins; Chapman &Hall: London, 1994.

302 Chem. Mater., Vol. 18, No. 2, 2006 Liang et al.

regime intermediate between that of epoxy resins and thehazardous/difficult to handle polyimide or BMI resins.Cyanate ester resins are easy to process in a manner similarto epoxy resins. Cyanate esters undergo thermal or catalyticcyclotrimerization to form triazine rings during curing. Weare studying cyanate esters as precursors to carbon-carbonmaterials because they may be cured without outgassing andthey form dense thermally stable resins. However, cyanateesters tend to be brittle and have reduced impact resistancelike most highly cross-linked thermosets.

The objective of this work is to incorporate nanosizedsilica phases into carbon-carbon materials. First, the natureof the dispersion after curing the resin needs to be under-stood. Can such nanodispersed phases be maintained uponpyrolysis of the cured cyanate ester resin/carbon clothcomposite and the subsequent densification processes? Ifnanodispersed phases occur, as envisioned, will they functionto form O2 diffusion barriers and to serve as char-formingnuclei? The specific POSS compound, TriSilanolPhenyl-POSS,1, is used as the nanophase in this work. Willincorporating this nanophase improve some of the propertiesof the original cyanate ester resin matrix? Can molecularlevel dispersions of1 within the cured cyanate ester resinbe achieved or will aggregation/phase separation occur? Ifmolecular level dispersion of1 is achieved within the curedresin, this may lead to carbon-carbon composites withdifferent properties than if phase-separated aggregates form.When POSS chemicals are blended with liquid resins, theymay or may not be able to dissolve in the resin. Even whenthey are soluble, they may phase separate to varying degreesduring resin curing as the entropy of mixing decreases.

This paper reports the blending of TriSilanolPhenyl-POSS,1, into cyanate ester resin (PT-15, Lonza Corp.), followedby thermal curing to determine the nature of cyanate ester/nanocomposites formed. POSS-1 was selected because thethree hydroxyl groups might aid solubility into PT-15 andwe thought they might react with cyanate ester functions athigh temperatures to aid dispersion.28,29 Cyanate ester/1composites with compositions (w/w) of 99/1, 97/3, 95/5, 90/10, and 85/15 were made and examined to determine theextent of phase separation of POSS into particle aggregatesversus the extent of compatible molecular dispersion. Theeffect of introducing POSS-1 into this cyanate ester polymersystem on dynamic mechanical properties and the thermaldimensional stability was also evaluated. Further work isunderway to study these PT-15/1 systems as the bondingmatrix for carbon-carbon composite materials.

Experimental Section

Materials. The phenolic-derived cyanate ester resin, PT-15, usedin this work was supplied by Lonza Inc. PT-15 is composed ofcyanate esters derived from a bisphenol-F mixture of2-4 withsome larger molecules such as5 included. Excess ClCN or BrCNconverted all phenolic hydroxyls to OCN functions in this bisphe-nol-F phenolic precursor mixture. The PT-15 monomer mixture,

represented by the single generalized structure6, is a multifunc-tional, low-viscosity (35 cps at 80°C) liquid cyanate ester resin.PT-15 is cured via a thermally driven cyclotrimerization to formtriazine rings, each of which serves as a cross-linking site. Thisreaction can take place readily in the absence of catalyst attemperatures above 165°C.

Multifunctional POSS, TriSilanolPhenyl-POSS, 1(C42H38O12Si7, MW ) 931.34 g/mol), with three SiOH groups, waspurchased from Hybrid Plastics Inc.

Preparation of Composites. PT-15/POSS composites wereprepared by a direct blending process. PT-15 (9.9, 9.7, 9.5, 9.0, or8.5 g, respectively) was heated to 120°C (η ) 8 cps) and held for10 min while stirring magnetically. Then, TriSilanolPhenyl-POSS-1(0.1, 0.3, 0.5, 1.0, or 1.5 g, respectively) was added as a powderinto the low-viscosity liquid resin. These mixtures (total weight 10g) were magnetically stirred for 50 min, during which time noviscosity increase occurred. POSS-1 appeared to completely dissolvein each case into the clear, transparent PT-15 liquid. These solutionswere placed into a mold without degassing and each sample wasoven-cured. The cure protocol was as follows: heat to 188°C andhold for 120 min, and then the temperature was ramped to 250°Cat 5°C/min and held at 250°C for 180 min. Each sample was thenpostcured at 300°C for 30 min.

Measurements.The Fourier transform infrared spectroscopy(FTIR) measurements were conducted on a MIDAC Co. instrumentat room temperature with 4 cm-1 spectral resolution by averaging1024 scans. Cured samples were ground with KBr and pressed intopellets for measurement. Granulated specimens (∼1.0 g) of everycomposite were immersed into a large excess of THF at roomtemperature for 6 months to see if any POSS could be extractedby THF. Soxhlet extraction for 24 h in THF was also performedfor each ground sample. Extraction of1 would indicate that theextracted POSS was not chemically bonded into the resin matrix.After extraction, the THF solutions were concentrated and coatedonto KBr plates and THF was removed for FTIR analysis of residue.

X-ray diffraction (XRD) measurements were performed toexamine potential POSS alteration of solid-state polymer micro-structure in the PT-15/1 composites and to see if ordered (crystal-line) POSS-1 aggregates form by phase separation. Samples wereexamined using a Philips XPERT model X-ray diffractometer.Philips Analytical software and Cu KR radiation (40 kV, 45 mA)were employed. Scans were taken over the 2θ range of 1-30° witha step size of 0.03° at 1 s/step.

Small-angle neutron scattering (SANS) experiments were per-formed at room temperature over theq range from 0.0038 to 0.2838Å-1 using the 30 m SANS instrument (NG7) at the NationalInstitute of Standards and Technology (NIST) Center for NeutronResearch (NCNR). The incident neutron wavelength wasλ ) 6 Åand the intensity at the sample was 17679 counts/s. The cross-sectional area of the beam was 50 mm× 50 mm and the entirecross section of the beam passed through the samples. The samples(100 mm diameters) were all solids. The scattered neutron intensitieswere corrected for background to yield the scattered intensities,I(q), as a function of the wave vector,q, whereq ) (4π/λ)sin(θ/2)

(28) Grigat, E.; Putter, R.Angew. Chem., Int. Ed. Engl.1967, 6, 206.(29) Shimp, D. A.; Christenson, J. R.; Ising, S. J.AroCy Cyanate Ester

Resins: Chemistry, Properties and Applications; Rhone-PoulencInc.: Louisville, 1991.

Cyanate Ester/POSS Nanocomposites Chem. Mater., Vol. 18, No. 2, 2006303

(θ is the scattering angle). IGOR software from WaveMetrics wasemployed for SANS data reduction.

A JSM-6500F field emission scanning electron microscope(FESEM) (JEOL USA Inc.) with an attached X-ray energydispersive spectrometer (X-EDS) was used to obtain elementalcompositions of the aggregated particles in PT-15/POSS compositesand to examine the fracture surfaces after three-point bendingflexural tests of neat PT-15 and PT-15/POSS composites. Thesamples for SEM were coated with gold before SEM measurements.The electron beam spot size used in X-EDS is about 5 nm indiameter.

Transmission electron microscopy (TEM) was used to identifyand characterize phase separation of1 in these composites. A JEOLJEM-2010 analytical transmission electron microscope (JEOL USAInc.), located at Texas A&M University, operating at 200 kV witha measured point-to-point resolution of 0.23 nm, was used tocharacterize the phase morphology in the PT-15/POSS composites.TEM samples were ultramicrotomed to∼60-80 nm thicknessesand mounted on carbon-coated Cu TEM grids. The JEM-2010instrument was equipped with an energy dispersive X-ray spec-troscopy (X-EDS) system with an electron beam spot size of∼2 nm.

The dynamic storage moduliE′ and loss factors (tanδ) weredetermined in the bending mode using a Rheometrics ScientificModel MK III DMTA instrument. A dual-level bending mode wasemployed. Small amplitude bending oscillations (both 1 and 10Hz) were used at a gap setting of 8.00 mm. Measurements werecarried out from 35 to 350°C on each composite. PT-15/POSS

test samples were approximately 3.0-4.0 mm thick, 4.5-5.5 wide,and 38 mm long.

Flexural strengths and flexural moduli were determined by three-point bending according to ASTM D-790-92 using a Zwickmaterials testing machine (model 1435) on 38 mm (length)× 10mm (width) × 3-4 mm (thick) specimens. The flexural strength(FS) value was calculated at specimen failure according to eq 1:

whereP is the breaking force of the specimen,L the support span,W the width, andt the thickness. The flexural modulus (FM) wascalculated from the tangent to the steepest initial straight-line portionof the load-deflection curve and using eq 2:

whereL is the support span,M the tangent of the initial straight-line portion of the load-deflection curve,W the width, andt thethickness.

Results and Discussion

Reaction of PT-15 and TriSilanolPhenyl-POSS.Themalcuring of PT-15 generates solid, cross-linked resins viacyclotrimerization of the OCN functions. This process forms

Scheme 1. Cross-linked Network Formation through Triazine Ring Formation

FS) 3PL

2Wt2(1)

FM ) L3M

4Wt3(2)

304 Chem. Mater., Vol. 18, No. 2, 2006 Liang et al.

trisubstituted triazine rings. The triazine rings serve asthermally stable cross-linking hubs throughout the resin. Inthe presence of1, what reactions occur between cyanateesters and1? Cyanate esters form addition compounds withphenols, alcohols, amines, and most other labile hydrogencompounds upon heating or in the presence of base cata-lysts.28 Amines add across cyanate esters at room tempera-ture,30 but phenol requires higher temperatures (100°C) forrapid reaction.31 These addition products are known to beimportant catalytic intermediates during triazine ring forma-tion.29 Thus, the silanol hydroxyl groups of1 can also addacross the CN triple bond of the cyanate ester groups ofPT-1529 as shown in Scheme 1 but they are less nucleophilicthan phenol and require a higher temperature.

FTIR confirmed a reaction proceeded during curingbetween the silanol hydroxyl groups of1 and OCN functionsof PT-15. Figure 1 shows the FTIR spectra for pure1, neatcured PT-15, and its PT-15/1 (85/15 w/w) composite.POSS-1 exhibits a very broad band centered at∼3248 cm-1,assigned to the stretching vibration of the hydrogen-bondedhydroxyl groups (O3Si-OH) of 1. Aromatic C-H stretchingappears at∼3069 cm-1. A strong Si-O-Si symmetricstretching peak appears at∼1100 cm-1 together with a sharpSi-O stretching vibration at∼1028 cm-1 for SiOH groups.32

In the PT-15/1 (85/15 w/w) composite, the stretching of theSi-OH groups is significantly decreased and the hydrogen-bonded hydroxyl group stretching peaks became indiscern-ible. The Si-O-Si symmetric stretching peak at∼1100cm-1 was present. Moreover, a new N-H in-plane bendingpeak appeared at∼1570 cm-1.33 This peak together withthe tris-oxygen-substituted triazine ring breathing vibrationat ∼1560 cm-1 appear (Figure 2).

Figures 3 and 4 show FTIR spectra of neat PT-15 and aPT-15/1 (85/15 w/w) composite, respectively, after incre-mental stepwise curing at different temperatures. The gradualdisappearance of-OCN peaks at 2240 and 2270 cm-1 andthe appearances of absorptions at 1365 and 1565 cm-1 fortriazine rings were observed. The intensity ratio of OCN(IOCN, at 2240 and 2270 cm-1) to the phenyl ring symmetric

breathing vibration (Iphenyl, at 1500 cm-1) decreases morequickly for PT-15/1 composite than for pure PT-15 whenheld at 120 or 188°C. The-OCN functions in PT-15 caneither cyclotrimerize to form triazine rings or react by addingSiOH functional groups of1 across the triple bond. Sincephenyl rings do not participate in the reaction of triazinering formation, the more rapid reduction of thisIOCN/Iphenyl

ratio in the PT-15/1 solutions must result from the additionaladdition reaction of SiOH across the-OCN functions. Thisaddition is illustrated in Scheme 1.

Further evidence for this reaction is the appearance of ahydrogen-bonded imine stretching band in the FTIR spectra.Figure 5 shows FTIR spectra from 1630 to 1530 cm-1 ofneat PT-15 and of the PT-15/1 composite after each washeated at 120°C for 1 h. A new absorption appeared at about1595 cm-1, which is due to the formation of the new CdNbonds by this SiOH addition in the PT-15/1 solution. Theabsorption of iminocarbonate groups, (RO)2CdNH, usuallyappears at 1640 cm-1.31 However, the (tSiO)(RO)CdNHfunctions generated at 120°C in the liquid state would behydrogen-bonded to other CdNH functions or to other SiOH

(30) Liang, K.; Toghiani, H.; Li, G.; Pittman, C. P.J. Polym. Sci., Part A:Polym. Chem.2005, 43, 3887-3898.

(31) Iijima, T.; Katsurayama, S.; Fukuda, W.; Tomoi, M.J. Appl. Polym.Sci.2000, 76, 208-219.

(32) Liu, H.; Zheng, S.; Nie, K.Macromolecules2005, 38, 5088-5097.(33) Choi, J.; Harcup, J.; Yee, A. F.; Zhu, Q.; Laine, R. M.J. Am. Chem.

Soc.2001, 123, 11420-11430.

Figure 1. FTIR spectra of TriSilanolPhenyl-POSS,1, neat cured PT-15,and a cured PT-(15/1 85/15) w/w composite.

Figure 2. FTIR spectra of TriSilanolPhenyl-POSS,1, neat cured PT-15,and its PT-15/1 (85/15 w/w) composite from 1850 to 950 cm-1.

Figure 3. FTIR spectra from 2500 to 1000 cm-1 of neat PT-15 cured todifferent stages.

Figure 4. FTIR spectra from 2500 to 1000 cm-1 of PT-15/1 (85/15 w/w)composite cured to different stages.

Cyanate Ester/POSS Nanocomposites Chem. Mater., Vol. 18, No. 2, 2006305

groups of1 prior to extensive resin curing. This would shiftthis absorption to a lower frequency, which we observe at1595 cm-1. This is similar to the 35 cm-1 lowering of theIR absorption of the CdN bond in-ArO(OdP)CdNH whenit becomes hydrogen-bonded.34

All samples were ground into powders and immersed inTHF for 6 months to dissolve and extract oligomeric andlinear polymeric components and any unreacted1. Eachground sample was also Soxhlet-extracted for 24 h in THF.1 was not detected in the residues isolated by extracting thesepowdered samples of the PT-15/1 99/1, 97/3, and 95/5composites and then evaporating the THF after filtering. TheFTIR spectra of 1 has a strong Si-O-Si symmetricstretching peak at∼1100 cm-1 and a sharp Si-O stretchingvibration (SiOH groups) at∼1030 cm-1. These bands werenot present in the extracted residues from the PT-15/1(99/1, 97/3, or 95/5 wt/ wt) composites. Furthermore, onlytraces of a Si-O-Si symmetric stretching peak at∼1100cm-1 was detected in the residues isolated by extractingpowdered samples of the PT-15/1 90/10 and 85/15 compos-ites with THF. These FTIR results are consistent withchemical incorporation of1 into the matrix as proposed inScheme 1. Silanol functions add across the cyanate estertriple bond when the temperature is raised to 120°C andabove. Triazine ring formation proceeds above 150-160°C.Scheme 1 presents a reasonable representation of the resin’sstructure. However, further increases in temperature to 250-300 °C must induce other changes in the structure. Itincreases cross-link density and likely causes further reactionsat the imine functions in the composite.

Blending POSS-1 with PT-15.POSS-1 exhibited goodsolubility in PT-15 at 120°C and PT-15/1 99/1, 97/3, 95/5,and 90/10 liquid mixtures were transparent. The goodsolubility was due to the formation of specific intermolecularinteractions (e.g., hydrogen bonding,-Si-OH‚‚‚NtC-O-)and phenyl ringπ-interactions between1 and PT-15. Afterthe temperature was held at 188°C for 120 min, all of thesepartially cured mixtures were still transparent. When the oventemperature was ramped to 250°C at 5°C/min and held at250°C for 180 min, PT-15/1 99/1, 97/3, and 95/5 compositesremained transparent. However, the PT-15/1 90/10 compositebecame translucent, suggesting a dispersed phase formedwith phase domain dimensions equal to or greater than visiblelight wavelengths. X-ray diffraction measurements were

performed to try to detect (1) any alteration of solid stateorder in the PT-15 cured resin’s microstructure and (2) tosee if ordered POSS aggregates form by phase separation.

Wide-Angle X-ray Diffraction (WAXD). The WAXDpatterns for the cured PT-15 and the PT-15/1 compositeswith w/w compositions of 99/1, 97/3, 95/5, and 90/10,respectively, are displayed in Figure 6. Figure 7 shows thewide-angle X-ray diffraction pattern for as-received1. Onebroad peak was observed at 2θ ≈ 20° in the WAXD pat-tern of neat PT-15. This broad peak is attributed to curedamorphous PT-15. The XRD patterns for the PT-15/1(99/1, 97/3, and 95/5) samples were essentially the same asthat for the cured pure PT-15 sample. Thus, after curing, nospecific evidence was found for aggregation of POSS-1 intoparticles having a regular crystalline structural organizationat loadings of1 up to 5 wt %. In Figure 3, however, a smallpeak at 2θ ≈ 7.5° (equivalent to an interplanar spacing of1.18 nm) is observed for the PT-15/1 (90/10) composite. Thispeak corresponds to the distance between the TriSilanol-Phenyl-POSS moieties found in the pure solid sample of1.Thus, some aggregation of1 into ordered particles must haveoccurred in this highly loaded PT-15/1 (90/10) sample.

The solubility of 1 decreases during the cure, due to asmaller contribution by the positive entropy of mixing.However, substantial reaction of1 with the resin has alsooccurred. When 10 wt % of1 is present, aggregation andparticle formation occurs due to phase separation whichbegins to occur at some point during the curing. The amountof solid 1 present and its degree of ordering in this sampleis just sufficient for its XRD peak at 7.5° (2θ) to be observed.Some regular structure occurs within some POSS aggregateswhich are present in these phase-separated PT-15/1 (90/10)composites. The absence of any observable peak at 7.5° (2θ)(34) Lin, C. H.Polymer2004, 7911-7926.

Figure 5. FTIR spectra from 1630 to 1530 cm-1 of neat PT-15 and itsPT-15/1 (85/15 w/w) composite after holding both for 1 h at 120°C.

Figure 6. X-ray diffraction patterns of neat cured PT-15 resin and itsPT-15/1 99/1, 97/3, 95/5, and 90/10 w/w composites.

Figure 7. X-ray diffraction pattern of TriSilanolPhenyl-POSS,1.

306 Chem. Mater., Vol. 18, No. 2, 2006 Liang et al.

suggests that there is very little or no aggregation of1 inPT-15/1 (99/1, 97/3, and 95/5) composites. If aggregationoccurs, it incorporates significant amounts of cyanate esterresin. In other words, molecular dispersion or formation ofsmall unstructured diffuse aggregates of1, or a combinationof both, was achieved in these 99/1, 97/3, and 95/5composites (see SANS studies).

The X-ray diffraction pattern of the PT-15/1 (99/1)composite overlaps exactly that of neat PT-15. Thus, thecrystallographic structure of PT-15 does not change ap-preciably by adding only 1 wt % of1. On the other hand,the intensity of the peak at 20° (2θ) decreases and graduallyshifts to a slightly smaller angle (about 19° (2θ)) upon addingmore 1 into the PT-15 resin matrix. This peak shiftcorresponds to a progressive microstructural modification inthe PT-15 resin matrix structure as the wt % of1 increasesfrom 3 to 5 to 10%. This changes the average layer spacingof PT-15 resin from 0.45 to 0.47 nm.

Small-Angle Neutron Scattering (SANS).Small-angleneutron scattering (SANS) is an excellent tool to analyzethe size and dispersion of particles in a continuous medium.35

The detection range of the NG7 30 m SANS instrument atthe National Institute of Standards and Technology (NIST)Center for Neutron Research (NCNR) approaches 1 nm. Thisis about the size of a single POSS particle size∼1 nm indiameter if one includes the first atoms attached to eachcorner Si.

In a neutron scattering, the scattered intensity depends onthe scattering length of the components forming the system,their concentration, and their relative positions in thesample.35c,36 The coherent contribution to the scatteredintensity,I(q), from a two-component system with a particlevolume fraction of ø and a single particle volume ofVp isproportional to the differential cross section per unit volumeof the sample, dΣ/dΩ (cm-1).37 This term is given by

with

whereb1 andb2 are the scattering lengths of components 1and 2. TheV1 andV2 terms are the monomer volume of thematrix polymer (component 1) and the molecular volumeof the second component, respectively. This equation ex-presses the total coherent scattering in terms of a singleparticle form factor,F(q), and the structure factor,S(q), which

accounts for interparticle interference. The scattering contrastin PT-15/1 composites comes from the difference betweenthe scattering lengths of the polymer matrix and the POSS-1particles. Structure factor,S(q), is the Fourier transform ofthe correlation function of the mass center of the POSSparticles. Thus, the SANS measurements give the Fouriertransform of the POSS density.

Figure 8 displays sample SANS scattering profiles ofPT-15/1 95/5 and 90/10 w/w composites. The slopes of theSANS profiles of 95/5 and 90/10 composites were notconstant over the experimentalq range. The scatteringintensity of the 90/10 sample was higher than that from the95/5 sample. The scattering curves were analyzed accordingto the ranges ofq values, starting from 0.1 Å-1, whichcorresponds to intraparticle dimensions (a few nanometers),and ends at 0.0038 Å-1, which corresponds to the large-scale organization of about 165 nm.

At q values between 0.1 and 0.04 Å-1, the scatteringintensities of the 95/5 sample display a plateau. The plateau’sintensity indicates there is little density fluctuation between63 and 160 Å. Thus, the number density of particles (from63 to 160 Å diameters) is similar over this length scale.38

The scattering intensities from the 90/10 sample follow aq-5/3 power law, where an exponent of-5/3 is indicativefor swollen single chains.39

Between 0.04 and 0.01 Å-1, the scattering intensities ofboth samples (5 and 10 wt %1) follow q-1 power law. Anexponent with a value of-1 corresponds to the scatteringfrom rodlike structures39 or aggregates that are very tenuousor even stringy.38 The q-1 power law over 0.04 and 0.01Å-1 implies that some molecules of1 aggregate to formnano-rod-like domain diameters from 160 to 600 Å.

Between 0.01 and 0.007 Å-1, the scattering intensityexponents of both the 5 and 10 wt % POSS-1 samples change

(35) (a) Yoo, J. N.; Sperling, L. H.; Glinka, C. J.; Klein, A.Macromolecules1990, 23, 3962-3967. (b) Yoo, J. N.; Sperling, L. H.; Glinka, C. J.;Klein, A. Macromolecules1991, 24, 2868-2876. (c) Hanley, H. J.M.; Muzny, C. D.; Ho, D. L.; Glinka, C. J.; Manias, E.Int. J.Thermophys.2001, 22, 1435-1448. (d) Ho, D. L.; Briber, R. M.;Glinka, C. J.Chem. Mater.2001, 13, 1923-1931. (e) Ho, D. L.;Glinka, C. J.Chem. Mater.2003, 15, 1309-1312. (f) Hanley, H. J.M.; Muzny, C. D.; Ho, D. L.; Glinka, C. J.Langmuir2003, 19, 5575-5580.

(36) (a) Chen, S. H.; Lin, T. S.Methods in Experimental Physics,Vol. 23,Part B. Neutron Scattering; Price, D. L., Skold, K., Eds.; AcademicPress: London, 1987. (b) Higgins, J. S.; Benoit, H. C.Polymers andNeutron Scattering; Clarendon Press: Oxford, 1994. (c) Lin, Y.;Alexandridis, P.J. Phys. Chem. B2002, 106, 12124-12132.

(37) Berriot, J.; Montes, H.; Martin, F.; Mauger, M.; Pyckhout-Hintzen,W.; Meier, G.; Frielinghaus, H.Polymer2003, 44, 4909-4919.

(38) Pignon, F.; Magnin, a.; Piau, J.-M.; Cabane, B.; Linder, P.; Diat, O.Phys. ReV. 1997, 56, 3281-3289.

(39) Radulescu, A.; Mathers, R. T.; Coates, G. W.; Richter, D.; Fetters, L.J. Macromolecules2004, 37, 6962-6971.

dΣ/dΩ ) b2øVpF(q)S(q)

b ) (b1 - b2(V1/V2))

Figure 8. SANS profiles of PT-15/1 95/5 and 90/10 w/w composites.

Figure 9. Schematic representation of formation of POSS-enrichedaggregated regions.

Cyanate Ester/POSS Nanocomposites Chem. Mater., Vol. 18, No. 2, 2006307

from -1 to -2.5, indicating some further aggregation toclusters 600-900 Å in size. Finally, atq values below 0.007Å-1, the scattering intensity exponents from both sampleschange from-2.5 to-4.5, indicating that aggregates withsizes>900 Å are present in both samples.

A slope of -2 indicates scattering from platelet shapes,while a slope of-4.5 is consistent with scattering fromapproximately spherical or from fractal-shaped clusters.40 Theexponent change of both 5 and 10 wt % samples can beexplained by a two-step aggregation process: first, POSSmolecules associate to form rodlike subunits with sizesranging from 160 to 600 Å. Then these subunits associateto form larger aggregates with a more spherical shape (Figure9). These aggregates do not have sufficient order to exhibitWAXD peaks in the 1, 3, or 5 wt %1 samples and onlyregister an extremely weak peak for the 10 wt % sample.Coughlin demonstrated that pendent POSS units present inethylene-norbornylene-substituted POSS copolymers werefound to aggregate and form elongated nonspherical latticesseparate from crystalline polyethylene lattice regions.41

Scanning Electron Microscopy (SEM).Figure 10 dis-plays two sample SEM micrographs of the translucentPT-15/1 90/10 composite. Table 1 summarizes the corre-sponding X-ray EDS elemental analysis of this compositeat locations shown in Figure 10. Clearly, some POSS-1 unitshave aggregated into approximately spherical POSS-richparticles at this high POSS-1 loading. The numbers shownon the SEM micrographs refer to the EDS spectrum numberrepresented in Figure 10. The heads of the arrows locatewhere the X-ray beams impinged on the sample. The spotsize of the impinged X-ray beams were small (∼5 nmdiameter) relative to the feature sizes observed in the SEMs.All of the regions studied exhibited the presence of1 asindicated by the value of the silicon content. Pure1(C42H38O12Si7) contains 21.1 wt % and 7.07 atom % Si. Thehighest Si content exhibited by EDS spectrum 2 (Figure 10a.)contained only 12.13 wt % (2.56 atom %) Si, confirming

that the POSS-rich phase-separated regions contain substan-tial amounts of PT-15 resin. This is further confirmed bytheir substantial nitrogen contents (from 9.89 to 13.16 wt %in Table 1.). Also, the weight percentages of POSS are dif-ferent from point to point within the same POSS-rich do-main. These findings are consistent with the very small XRDpeak at 7.5° (2θ) in this 10 wt %1 composite (Figure 6).The large aggregates cause the 10 wt %1 sample to betranslucent. The substantial PT-15 content within theseaggregates is consistent with the clustering route proposedin Figure 9.

(40) Buckley, C. E.; Birnbaum, H. K.; Lin, J. S.; Spooner, S.; Bellmann,D.; Staron, P.; Udovic, T. J.; Hollar, E.J. Appl. Crystallogr.2001,34, 119-129.

(41) Zheng, L.; Waddon, A. J.; Farris, R. J.; Coughlin, E. B.Macromol-ecules2002, 35, 2375-2379.

Figure 10. Two SEM micrographs of the PT-15/1 90/10 composite.

Table 1. X-ray EDS Spectra Data of PT-15/1 90/10 Composite(Figure 10)

spectrum element weight % atomic %POSS

weight %

1a C 68.85 74.10N 11.72 10.81O 17.68 14.29Si 1.74 0.80 8.24

2a C 65.56 71.27N 13.16 12.27O 18.71 15.27Si 2.56 1.18 12.13

3a C 70.22 74.92N 11.45 10.48O 18.08 14.48Si 0.25 0.11 1.18

4a C 66.72 72.31N 12.93 12.01O 17.84 14.51Si 2.52 1.17 11.94

5a C 68.79 74.15N 12.13 11.21O 16.79 13.59Si 2.29 1.05 10.85

1b C 71.01 76.18N 9.89 9.10O 17.18 13.84Si 1.92 0.88 9.10

2b C 69.40 74.51N 12.08 11.12O 16.91 13.63Si 1.61 0.74 7.63

3b C 68.90 73.88N 12.83 11.79O 17.18 13.83Si 1.09 0.50 5.16

308 Chem. Mater., Vol. 18, No. 2, 2006 Liang et al.

Transmission Electron Microscopy. PT-15/1 99/1,97/3, and 95/5 composites were transparent. The 97/3 and95/5 composite samples were selected for TEM measure-ments. Irregular regions appear in the TEMs of the nano-composites (Figures 11 and 12), ranging from 200 nm to afew nanometers in diameter. The composition of theseregions was studied by X-EDS and compared to the com-position of what appears to be the continuous matrix regions.

X-EDS measurements detected substantial amounts ofsilicon in the continuous phase matrix of both the 97/3 and95/5 samples. No silicon was found in cured samples of purePT-15 resin. The regions of the 97/3 composite, whichvisually might appear to be particles in TEM micrographs,exhibit almost the same Si/O/C composition ratio as thecontinuous phase of the matrix (Figure 11). Thus, thesedarker regions may not really be POSS-rich aggregates. The“particle-like” inclusion or artifact in the 97/3 sample towhich the arrow points (in Figure 11) has dimensions ofabout 50× 25 nm. This feature exhibits the same Si/O/Ccomposition within experimental error as the matrix region.

A TEM of the 95/5 composite is shown in Figure 12. With5 wt % of POSS-1 present, the observed particle-likeinclusions were found to have a greater Si content than thesurrounding matrix. The Si/O/C compositions varied fordifferent particle (aggregate) regions in the 95/5 sample, butall of these domains contained a higher Si content than thecontinuous phase (Figure 12). Since the PT-15 resin itselfdoes not contain any Si, the Si detected in the continuousmatrix must be due to molecular level incorporation of1within the resin. This is consistent with a curing process thatproceeds through the general chemical pathway shown inScheme 1. The domains observed with higher Si contentsthan the matrix are best described as POSS-enriched ag-gregates because they do not contain as much Si as1. Thus,

the appropriate interpretation is that some1 has beendispersed molecularly within the matrix and the remainderforms aggregates which also incorporate various substantialamounts of resin. This means highly crystalline structuresdo not form, which agrees with the absence of XRD peaksfor 1 in the 95/ 5 sample (Figure 6).

Various particles observed by TEM (Figures 11 and 12)could also have small amounts of matrix either above orbelow their boundaries, which is detected by X-EDS. Thiswould also lead to different elemental ratios. However, theX-EDS spot size is very small (10 nm2), giving confidenceto the conclusion that the matrix is not being sampled withinthe plane of the micrograph because the particle areas studiedby X-EDS are far larger than the X-ray beam. Thus, theTEM/X-EDS results augment the results from XRD andSEM/X-EDS. Each POSS-1 core is surrounded by sevenphenyl groups and contains three silanol (Si-OH) groups.Many POSS macromers are chemically bonded into thePT-15 resin. Aggregated regions, which also contain cyanateester matrix components, are formed. These regions involve1 that must be incorporated into the matrix since1 cannotbe extracted from these samples.

The TEM/X-EDS observations suggest that most POSSunits are well-dispersed in the PT-15/1 97/3 and 95/5samples. Two factors lead to the good dispersion of1 inPT-15. First, 1 has a substantial solubility in uncuredPT-15. This likely results from hydrogen bonding betweenthe fairly acidic silanol OH groups and the nitrogen oroxygens of the cyanate ester functions. The seven phenylrings of 1 can also be solvated by aryl rings in PT-15.Second,1 is incorporated by chemical bonding into thedeveloping PT-15 resin before and during curing as thecomposite forms. Thus, when the solubility limit of1 wasreached during curing, some1 is already bound into cross-

Figure 11. Progressive magnification TEM micrographs of the PT-15/POSS-1 (97/3) composite showing molecular dispersion of1 in PT-15. Si frommolecules of1 is dispersed in the resin matrix. The dark particle has the same Si/O/C composition as the resin matrix.

Cyanate Ester/POSS Nanocomposites Chem. Mater., Vol. 18, No. 2, 2006309

linked matrix regions and no longer free to self-aggregate.This limits the extent of phase separation. Furthermore, theconcentration of unreacted1 in the resin decreases as it reactswith the resin. As curing proceeds, the viscosity increases.This decreases the diffusion rates of1, thereby slowingaggregation. As cross-linking occurs, further aggregation ofresin-bound1 becomes more difficult.

In the 95/5 composite only small-sized (<200 nm) POSS-1-rich regions were formed. Some POSS-1 could react withsmaller cyanate ester molecules (monomers and oligomers)which are not yet cross-linked. A portion of these mayaggregate, prior to further polymerization, which would thengenerate aggregates rich in POSS-1 that contain cyanate esterwhich is chemically linked with the continuous matrixregions.

Mechanical Properties. DMTA Studies.The bendingstorage modulus,E′, versus temperature curves at 1 Hz forneat PT-15 and each PT-15/1 composite are shown in Figure13. TheE′ (both above and belowTg) values of the 99/1,97/3, and 95/5 composites are higher than those of the purePT-15 over the entire 35-350°C temperature range. Thesethree composites also displayTg values∼15 to 25°C higherthan theTg of the pure PT-15 prepared under identical

conditions. Importantly, theirE′ values aboveTg are signifi-cantly higher than that of the cured PT-15 resin. As theamount of1 increases to 10 and 15 wt %, theTg values dropas do theE′ values atT > Tg. While PT-15/1 90/10 and85/15 composites exhibit higherE′ values than those of thepure PT-15 in the glassy region (T < Tg), the 15 wt %POSS-1 sample exhibited lowerE′ values in the rubberyregion (T > Tg). Furthermore, theTg values are lowered by25 and 35°C in the 10 and 15 wt % POSS composites,

Figure 12. Higher magnification TEM micrographs of the PT-15/POSS-1 (95/5) composite showing molecular dispersion of1 in PT-15. Si is dispersed inthe resin matrix. The particles have a higher Si content than the resin.

Figure 13. Bending modulus (E′) versus temperature curves at 1 Hz (fromDMTA) for neat PT-15 and its PT-15/1 99/1, 97/3, 95/5, 90/10, and 85/15w/w composites.

310 Chem. Mater., Vol. 18, No. 2, 2006 Liang et al.

respectively, versus pure PT-15. Thus, small amounts (e5wt %) of 1 incorporated into PT-15 can increase the storagemodulus in the glassy and rubbery regions and theTg valuesversus pure PT-15. As the amount of1 increases (10 and 15wt %), theTg andE′ (T > Tg) values drop below that of theparent cyanate ester resin.

The bending tanδ versus temperature curves obtained at1 Hz (from DMTA) for the pure PT-15 and the PT-15/1composites are shown in Figure 14. TheTg values for thePT-15 resin and its 99/1, 97/3, 95/5, 90/10, and 85/15composites, defined as the tanδ peak temperatures, are 305,325, 321, 313, 278, and 260°C, respectively. The 99/1sample has the highestTg among these samples. TheTg

values then decrease as the content of1 increases.

Two factors affect the viscoelastic properties. First,1 isrelatively rigid. Massive units of1 along chain segment unitswill tend to retard or restrict that segment’s motions inPT-15. However,1 may also increase free volume by notpacking as well overall with the cyanate ester resin segmentsof the cross-linked resin. The first factor increases both therigidity and theTg while the second adds free volume, lowersTg, and could allow increased segmental motion. A portionof the three available acidic silanol (Si-OH) groups of1can be chemically bound into cured PT-15 (Scheme 1).Incorporating1 may decrease the overall cross-linking den-sity since the volume of each POSS-1 molecule is substantial.This in turn could decreaseTg of the composite by loweringthe cross-link density per unit volume. When the reductionof segmental motion dominates, then the rigidity andTg ofthe composite will increase. However, at larger loadings of1 the trend reverses. A lower net cross-link density and/orincreased free volume is consistent with the increased flexuralstrength of the PT-15/1 85/15 w/w sample which couldundergo greater bending (see flexural properties). Changes

Figure 14. Bending tanδ versus temperature curves at 1 Hz (from DMTA)for neat PT-15 and its PT-15/1 99/1, 97/3, 95/5, 90/10, and 85/15 w/wcomposites.

Figure 15. Fracture surface SEM micrographs: (a) neat PT-15 and (b-d) PT-15/1 85/15 w/w composite.

Table 2. Flexural Strength (FS) and Flexural Moduli (FM) ofPT-15/POSS Compositesa

samples FS (MPa) FM (MPa)

PT-15 120.72 3337.78PT-15/POSS 99/1 121.18 3385.99PT-15/POSS 97/3 122.29 3832.38PT-15/POSS 95/5 123.54 4267.73PT-15/POSS 90/10 124.64 3634.95PT-15/POSS 85/15 145.43 3558.28

a Six samples of each were averaged.

Cyanate Ester/POSS Nanocomposites Chem. Mater., Vol. 18, No. 2, 2006311

in the extent of aggregation and detailed morphology alsooccur as the content of1 increases (g10 wt %). Finally, asthe loading of1 goes up, the remaining acidic SiOH functionsmay modify the degree of triazine formation.

Flexural Properties and Fracture Surfaces.The flexuralstrengths and moduli of the PT-15/1 composites are shownin Table 2. The flexural strengths of PT-15/1 samples areslightly higher than that of neat PT-15 resin and increase asthe loading of1 increases. Only the PT-15/1 85/15 sampleexhibits a sizable flexural strength improvement (20%). Also,the flexural moduli are higher than that of neat PT-15 resin.The flexural moduli increase as the content of1 increasesand then decrease as the wt % of1 increases to 10 and 15%.The PT-15/1 95/5 sample exhibits the highest value, 28%above that of the neat PT-15 resin. Therefore, the mechanicalproperties of PT-15 resin can be improved by incorporationof 1.

Fracture surface SEM micrographs of the neat PT-15 andthe PT-15/1 85/15 w/w composite after failure in the three-point bend flexural test are shown in Figure 15. Crackpropagation, as indicated by failure ridges, is straighter forthe neat PT-15 resin (Figure 15a). In contrast, these failureridges seem to interact with and terminate at POSS-richaggregates during crack propagation within the PT-15/1(85/15) composite (Figures 15b-d). Some blunting of thefailure process occurs. The aggregates also seem to act ascrack attractors for a propagating crack front. Intermittentinterruption of crack growth gives rise to a more complexcrack propagation path instead of a relatively straight/smoothfracture path.

Conclusions

The liquid PT-15 cyanate ester resin dissolved1. Aftercuring, FTIR, XRD, SANS, and X-EDS analyses indicatedthat a portion of1 was molecularly dispersed and bondedinto the continuous matrix phase, while some was presentin aggregates and/or particles containing both1 and cya-nate ester resins. At 1 and 3 wt % POSS loadings, appreci-able fractions of1 are molecularly dispersed. SANS andTEM also detected POSS-enriched nanophases (from a few

nanometers to about 200 nm) in the composites. Thesenanophases also incorporate cyanate ester. The size distribu-tion of these phases broadens and larger aggregates increas-ingly form as the loading of1 increases. At 15 wt %1, thedispersed particulate phase clearly influences the fracturemechanism. The storage moduli,E′, of PT-15/1 99/1, 97/3,and 95/5 composites are higher than those of the pure curedPT-15 over the entire temperature range 35-350 °C. TheE′ values for all the composites (except that when 15 wt %of POSS was present) are greater than that of the pure resinat T > Tg. Also, theTg values of the 1, 3, and 5 wt % POSScomposites are higher than that of neat PT-15. Small amounts(e5 wt %) of POSS can improve the storage modulus themost. The flexural strength and flexural modulus are higherthan those of neat PT-15 and larger amounts of1 (>3 wt%) are most beneficial.

The aggregation/phase separation process is quite complex.First, 1 chemically reacts with cyanate esters to form iminosiloxycarbonates. This improves dispersion. Continued curingleads to aggregation of1 which is bound to resin molecules.The aggregation process depends on the concentration of1used. Cross-linking eventually gels the system but as thetemperature increases to 250°C the fate of the iminosiloxycarbonate functions is unknown. Some formation ofSiOSi bonds from silanol groups of1 may also occur, bothintramolecularly and intermolecularly at 250-300 °C. Thelatter could create dimers or trimers of1 when there isenough freedom of motion or when molecules of1 are inclose proximity.

Acknowledgment. The authors thank Z. P. Luo at TexasA&M University for obtaining the TEMs shown here. Theauthors acknowledge the support from the National Institute ofStandards and Technology, U.S. Department of Commerce, forproviding the neutron research facilities and funding (Proposalnumber S14-13, U14-02, Submission ID: 3456). Partial supportwas provided by the Air Force Office of Scientific ResearchSTTR program (Contract No. F49620-02-c-0086 and Grant No.F49620-02-1-026-0).

CM051582S

312 Chem. Mater., Vol. 18, No. 2, 2006 Liang et al.